The contribution of four cortical areas (S1, S2, insular cortex and gyrus cinguli) to pain processing was assessed by functional magnetic resonance imaging (fMRI). Phasic (mechanical impact) and tonic stimuli (squeezing) were applied to the back of a finger, both at two different strengths. Stimuli were adjusted to inflict weak and strong pain sensations. It had been shown before that stronger noxious mechanical stimuli induce a weaker input from myelinated mechanoreceptors, but a more vigorous input from nociceptive primary afferents, and vice versa. Sizes of activation clusters and percent increase of the blood oxygenation level dependent (BOLD) signal during activation were compared in the areas of interest. Phasic stimulus patterns were more closely reflected in the time course of the MR signal in S1, S2 and the cingulate than tonic patterns, since the tonic stimuli tended to induce slow MR signal increase also during the resting periods which is in parallel to the persisting character of the tonic pain sensations. In S1 only the contralateral side was activated in most cases, and the more painful stimuli did not induce greater BOLD responses compared to the less painful stimuli in this area. Paradoxically, more painful stimuli produced smaller activation clusters in S1, both in tonic and phasic stimulus trials. In contralateral S2 more painful phasic stimuli induced significantly stronger BOLD responses than the weaker stimuli. The responses to tonic stimuli did not differentiate painfulness and were significantly smaller than the phasic. Activation clusters in this area were also smaller for tonic stimuli. In the gyrus cinguli more painful phasic stimuli induced stronger BOLD responses, but no difference was seen between tonic stimulation of different strength. Though the insular cortex was often bilaterally activated, no significant differences between stimulus quality or intensity were found. Our results provide evidence for a contribution of the S2 projection area and of the cingulate cortex to the processing of the intensity dimension of phasic mechanical pain. Such evidence was not found for the S1 area, which probably receives dominant input from non-nociceptive mechanoreceptors.